JP5200715B2 - Electric inertia control device of dynamometer system - Google Patents

Description

  The present invention relates to an electric inertia control device for a dynamometer system.

FIG. 6 shows a configuration diagram of a chassis dynamometer system known from Patent Document 1 and the like. Dy is a dynamometer, R is a roller connected to the dynamometer Dy, IV is an inverter, and RP is an electric inertia. Control circuit, TM is an axial torque meter, EC1 is an encoder for detecting the rotational speed of the dynamometer, EC2 is an encoder for detecting the rotational speed of the roller, and is detected by these axial torque meter TM and encoders EC1 and EC2. Each detection signal is input to the electric inertia control circuit RP to calculate a torque current command.
JP 2004-361255 A

As described in Patent Document 1, an electric inertia control method is employed in which a mechanical inertia component of a power measurement target is electrically compensated on the load side or drive side of the measurement system, and feedback of shaft torque detection is described. ing. In the control method as shown in Patent Document 1, the mechanical system model of the dynamometer is a two-inertia system having a resonance characteristic, but the resonance characteristic of the mechanical system is not considered in the electric inertia control. The dynamometer system also has a delay in detection of shaft torque and a delay in torque response of the inverter. Therefore, when trying to increase the electrical inertia control response, hunting and divergence caused by the resonance characteristics of the mechanical system are avoided. A stable phenomenon occurs, and stable control with high response is difficult.

  Therefore, an object of the present invention is to provide a highly responsive and stable electric inertia control device.

The present invention connects a roller and a dynamometer through a shaft, and includes an observer section, an inertia section, a speed control section, and an axial torque control section in an electric inertia control circuit, a dynamometer rotation signal, a shaft torque signal, and a roller rotation signal. Is used to calculate a torque current command, and electric inertia control is performed via an inverter according to the obtained torque current command.
The observer unit inputs a shaft torque detection value, a roller or dynamometer angular velocity detection value, calculates a roller surface driving force estimation value, and inputs a difference signal between the output of the observer and a running resistance value via the inertia unit. And a speed control controller for executing PID calculation with the roller or dynamometer angular velocity detection value, a deviation signal between the output of the speed controller and the shaft torque detection value, a dynamometer angular velocity detection value, and a roller angular velocity detection value The shaft torque control controller that generates torque current command by inputting the H is provided with the shaft torque control controller designed by the controller design method called H∞ control and the design method based on the ATR generalized plant model It is characterized by that.

  Further, according to the present invention, the observer unit for calculating the roller surface driving force estimation value is composed of a circuit designed by a controller design method called H∞ control and a design method based on an observer generalized plant model. It is characterized by that.

  As described above, according to the present invention, the ATR controller constituting the electric inertia control circuit is designed by the controller design method called the H∞ control and the design method, and the resonance characteristic of the mechanical system, the shaft torque detection characteristic, the dynamometer angular velocity detection It is configured in consideration of characteristics, roller angular velocity detection characteristics, and inverter response characteristics. Thereby, a torque current command in which each characteristic is taken into account can be obtained. Therefore, the resonance characteristics are suppressed, and the electric inertia control of the dynamometer system with higher response and stability is possible. Further, even when the running resistance value is added or when the set inertia amount is changed, it is possible to easily cope with it.

The electric inertia control device of the present invention is provided with an observer for calculating the roller surface driving force estimation value, an inertia part, an ASR (speed control) controller, and an ATR (shaft torque control) controller, and the ATR controller is an ATR generalized plant model. Based on the H∞ control, a controller design method called a design method is used. This ATR controller has a difference signal between the output signal of the ASR controller and the detected shaft torque value,
The dynamometer angular velocity detection value and the roller angular velocity detection value are input to generate a torque current command. Hereinafter, description will be made based on examples.

FIG. 1 is a block diagram of an electric inertia control circuit showing an embodiment of the present invention, which has an observer (roller surface driving force estimation) 101, an inertia part 102, an ASR controller 103, and an ATR controller 104, and is used for electric inertia control. The set inertia amount is EIC _- J.

In the present invention, the observer unit 101 and the ATR controller 104 in the electric inertia control circuit of such a chassis dynamometer system are designed by a controller design method called a H∞ control and a design method, and the ASR controller and the inertia unit are arranged. By combining, a moment of inertia equivalent to the vehicle body is designed. The observer 101 receives shaft torque detection and dynamometer angular velocity detection, and calculates a roller surface driving force estimation value. The roller surface driving force estimated value is calculated by the subtractor 105 as a difference from the running resistance value and input to the inertia part 102. The inertia part 102 sets and integrates the set inertia of 1 / EIC ₋ J. The integrated signal is input to the ASR controller 103. The ASR controller 103 also receives the detected value of the dynamometer angular velocity, executes the PID calculation, obtains the torque setting Tref, and outputs it to the subtraction unit 106. The subtraction unit 106 performs a difference calculation between the torque setting Tref and the shaft torque detection, and inputs the deviation signal to the ATR controller 104 as the shaft torque command SHT.e. The ATR controller 104 also receives a roller angular velocity detection w _- roller.det and a dynamometer angular velocity detection w _- dy.det, and calculates a torque current command based on these input signals.

FIG. 2 shows an example of a generalized plant model of an observer used in the present invention, and FIG. 3 shows a transfer function of a mechanical system model in the observer generalized plant model.
The generalized plant model of the observer shown in FIG. 2 receives the roller surface driving force w1, the dynamometer torque command w2, the shaft torque observation noise w3, and the dynamometer angular velocity observation noise w4 as disturbances. Is input to the controller 10. The disturbance signal assumes that the chassis dynamometer system is actually driven, and noise generated in the control loop at that time is a disturbance signal, and here, there are four disturbance signals w1 to w4. The controller 10 sets the parameters of the state equation of the observer, and executes a predetermined calculation for determining the parameters based on the algorithm so that the gain becomes small. In the generalized plant model, z1 to z4 are generated as control amounts.

  The input disturbances are weighted individually in weighting factor adding means 1 (Ow1 (s)) to 4 (Ow4 (s)) and 11 (Oz1 (s)) to 14 (Oz4 (s)), respectively. The desired characteristics can be obtained. That is, in the means 1, a certain constant is multiplied by the weight applied to the vehicle driving force, and it is output to the mechanical system model 20 (Omec (s)) as the rotational moment torque J1.T of the roller and to the subtracting means 7 Is done. In the means 2, a certain constant is multiplied by the weight applied to the torque command of the dynamometer, and the output is output to the mechanical system model 20 as J2.T. In the means 3, the weight applied to the detection error of the shaft torque is set to a characteristic such that the gain increases at a certain constant or high frequency, and is output to the adding means 5. In the means 4, the weight applied to the dynamometer angular velocity detection error is set to a certain constant or a characteristic such that the gain is increased in a high range, and is output to the adding means 6.

  8 (Otm (s)) is a torque meter characteristic generating means (torque meter characteristic model) for detecting the shaft torque. The shaft torque K12.T of the coupled shaft from the mechanical model 20 is input and added as a predetermined torque meter characteristic. Output to means 5. The adding means 5 adds the weighted shaft torque detection error signal and the torque meter characteristic to generate a shaft torque detection value, which is input to the controller 10 as the observation amount A. Reference numeral 9 (Oenc (s)) denotes an encoder characteristic generation means (encoder characteristic model), which inputs the dynamometer angular velocity J2.w from the machine model 20 to generate a predetermined encoder characteristic and outputs it to the adder 6. The adding unit 6 adds the weighted dynamometer angular velocity detection error signal and the encoder characteristic to generate a dynamometer angular velocity detection value, and outputs it to the controller 10 as an observation amount B. The controller 10 performs a predetermined calculation based on the input signal. The calculation signal is output to the subtracting means 7, subtracted from the rotational moment torque J1.T of the roller, and output to the means 11.

  The means 11 (Oz1 (s)) is a means for weighting the deviation value of the roller surface driving force estimation, and the weighted roller surface driving force is set to an integral characteristic or a predetermined characteristic such that the gain becomes low in a high range. The estimated value z1 is output. Means 12 (Oz2 (s)) is means for applying a weight to the shaft torque, and is a constant to which the shaft torque K12.T from the mechanical system model 20 is input or a predetermined value that increases the gain in the high range. The characteristic is output as a shaft torque signal z2. The means 13 (Oz3 (s)) is a means for applying a weight to the dynamometer angular velocity J2.w from the mechanical system model 20, and is weighted with a predetermined characteristic or a predetermined characteristic that increases the gain in a high frequency range. Output as a dynamometer angular velocity signal z3. Means 14 (Oz4 (s)) is a means for applying weight to roller angular velocity detection from the mechanical system model 20, and outputs it as a weighted roller angular velocity signal z4 as a characteristic that increases the gain in a certain constant or high frequency range.

An observer mechanical system model 20 shown in FIG. 3 represents a mechanical characteristic of a dynamometer by a transfer function, and is a two-inertia mechanical system model. The mechanical system model of this embodiment has J1.T and J2.T as inputs and J1.w, K12.T, and J2.w as outputs.
In the figure, 21 is a roller inertia moment element, and its output is outputted to the generalized plant as a roller angular velocity J1.w and also outputted to the subtracting means 26. A spring stiffness element 22 receives the difference signal between the dynamometer angular velocity and the roller angular velocity calculated by the subtracting means 26 and outputs it as a shaft torsion torque K12.T signal to the generalized plant. Output to. In the adding means 24, the rotational moment J1.T of the roller due to the vehicle driving force applied to the roller surface and the shaft torsion torque K12.T are added and input to the roller inertia moment element 21. Further, the subtracting means 25 obtains a difference signal between the input dynamometer torque signal J2.T and the shaft torsion torque K12.T and outputs it to the dynamometer inertia moment element 23. The angular velocity J2.w is calculated and output to the generalized plant and also output to the subtracting means 26.

FIG. 4 shows an example of a generalized plant model of the ATR unit 104 used in the present invention.
The ATR generalized plant model shown in FIG. 4 includes a roller surface driving force w11, an inverter torque control error w12, a shaft torque command w13, a roller angular velocity observation noise w14, a shaft torque observation noise w15, and a dynamometer angular velocity observation noise w16 as disturbances. Is input, and z11 to z15 are output as control amounts. Reference numeral 30 denotes an ATR controller to which observation quantities c _- in1, c _- in2, and c _- in3 are input. The ATR controller 30 sets the parameters of the state equation for shaft torque control, executes a predetermined calculation for determining the parameters based on the algorithm so as to reduce the gain, and generates a dynamometer torque command c _- out. To do. Here, c _- in1 is roller angular velocity detection, c _- in2 is shaft torque control deviation, and c _- in3 is dynamometer angular velocity detection. In the generalized plant model, z11 to z15 are generated as control amounts.

  The input disturbances are respectively weighted individually by weighting coefficient adding means 31 (Gw1 (s)) to 36 (Gw6 (s)) and 47 (Gz1 (s)) to 51 (Gz5 (s)). The desired characteristics can be obtained. In other words, the means 31 is weighted to the vehicle driving force and has a characteristic that increases the gain at a certain constant or in a high frequency range, and sets the rotational moment torque J1.T of the roller to the mechanical system model 60 (Gmec (s)). Entered. In the means 32, the torque current control error of the inverter is weighted, and a characteristic is set such that the gain increases at a certain constant or high frequency. In the means 33, the shaft torque command is weighted, and is given a certain constant or a characteristic such that the gain is increased in a high range, and is output to the subtracting unit 42. In the means 34, the roller angular velocity observation noise is weighted so as to have a characteristic that increases the gain in a certain constant or high frequency range. In the means 35, a constant weighted to the shaft torque observation noise or a characteristic such that the gain is increased in a high range is output to the adder 41. The means 36 weights the dynamometer angular velocity observation noise and outputs it to the adding unit 45 with a certain constant or a characteristic that increases the gain in a high frequency range.

Reference numeral 37 denotes an inverter characteristic model unit that generates an inverter response characteristic signal based on the output c _- out of the ATR controller 30, and adds the weighted signal in the means 32 to the adding unit 43 to obtain a dynamometer torque J2.T. Input to the mechanical system model 60.
Reference numeral 38 denotes a first encoder characteristic model. The roller angular velocity calculated by the mechanical system model 60 and the roller angular velocity observation noise w14 are weighted by the means 34, and the gain increases at a certain constant or in a high range. The sum (at the adder 44) with the roller angular velocity signal thus received is input to detect the roller angular velocity. This signal is input to the ATR controller 30 as an observation amount of the roller angular velocity detection c _- in1. Further, the roller angular velocity detection signal in the adding unit 44 is output to the means 51 and weighted, and is given a certain constant or a characteristic such that the gain is increased in a high range, and is set to a weighted roller angular velocity signal z15.

Reference numeral 39 denotes a torque meter characteristic model for detecting shaft torque, which is based on the sum signal (in the adder 41) of the shaft torque K12.T from the mechanical system model 60 and the shaft torque observation noise weighted by the means 35. A torque meter characteristic signal is generated and output to the subtracting unit 42. The subtractor 42 performs a difference calculation with the weighted shaft torque command by the means 33 and outputs the difference signal to the ATR controller 30 as a shaft torque control deviation c _- in2 and also to the means 46. In the means 46, a weight function having an integral characteristic is added to the input shaft torque control deviation so as to obtain a weight constant shaft torque control signal z13 having a certain constant or a characteristic such that the gain becomes low in a high range. Further, the shaft torque observation error signal obtained by the adder 41 is input to the means 48 and is weighted so as to have a characteristic that increases the gain in a certain constant or high frequency range, so that the weighted shaft torque signal z12. It becomes.

Reference numeral 40 denotes a second encoder characteristic model for detecting the dynamometer angular velocity, which is a sum signal of the dynamometer angular velocity J2.w from the mechanical system model 60 and the dynamometer angular velocity observation error weighted by the means 36 (in the adder 45). And an encoder characteristic signal is input to the ATR controller 30 as an observation amount of the dynamometer angular velocity c _- in3. Further, the sum signal in the adder 45 is weighted by the means 50, and is given a constant or a characteristic such that the gain becomes high in a high frequency, and becomes a weighted dynamometer angular velocity signal z14.

The ATR controller 30 sets the parameters of the state equation for shaft torque control based on the input observation quantities c _- in1, c _- in2, and c _- in3, and determines the parameters based on the algorithm so that the gain is reduced. Predetermined calculation is performed, and the calculated dynamometer torque command c _- out is generated and output to the inverter characteristic model unit 37 and output to the means 47. The means 47 weights the torque current command of the inverter and outputs it as a weighted torque current command signal z11 with a certain constant or a characteristic that increases the gain in a high range.
The mechanical system model used for the ATR generalized plant model is configured in the same manner as in FIG.

  Moreover, in the generalized plant model shown in FIG. 2 and FIG. 4, each input observation noise is arbitrarily selected. For example, when the dynamometer angular velocity observation noise is omitted, the related main stages or functional parts after the means for weighting the dynamometer angular velocity observation noise are removed.

According to this embodiment, the observer and the ATR controller constituting the electric inertia control circuit are designed by the controller design method called the H∞ control and the design method, and the resonance characteristic of the mechanical system, the shaft torque detection characteristic, the dynamometer Design that takes into account angular velocity detection characteristics, roller angular speed detection characteristics, and inverter response characteristics is possible, and a torque current command that takes these characteristics into consideration can be obtained. Therefore, resonance characteristics are suppressed, and electric inertia control of a chassis dynamometer system with higher response and stability is possible. Further, even when the running resistance value is added or when the set inertia amount is changed, it is possible to easily cope with it.

  FIG. 5 shows another example of an observer used in the present invention. In the above description, the generalized plant model is used that has been designed by the controller design method shown in FIG. 2, but it may be configured as shown in FIG. 5 to obtain the estimated roller surface driving force. That is, it is configured to estimate a dynamometer angular velocity detection value or a roller angular velocity detection value and a shaft torque detection value to estimate a driving force equivalent signal (roller surface driving force estimation) and output the estimated value to the subtraction unit 105. . Here, G (s) is an arbitrary transfer function having a relative degree of 1 or higher in the low-pass filter, Jroller is the moment of inertia of the roller, and s is a differential operation expressed by a Laplace operator s.

The block diagram of the electric inertia control apparatus which shows embodiment of this invention. The block diagram of an observer generalized plant model. Mechanical system model diagram. The block diagram of an ATR generalized plant model. The block diagram of a roller surface driving force estimation part (observer). The block diagram of an electric inertia control apparatus.

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 ... Observer 102 ... Inertial part 103 ... ASR part 104 ... ATR part 105, 106 ... Subtraction part 10 ... Controller 20,60 ... Mechanical system model 30 ... ATR controller Dy ... Dynamometer IV ... Inverter RP ... Electric inertia control circuit R ... Roller EC (EC1, EC2) ... Encoder TM ... Torque meter

Claims (2)

A roller and a dynamometer are connected via a shaft, and a dynamometer rotation signal, a shaft torque signal, and a roller rotation signal are input to an electric inertia control circuit having an observer unit, an inertia unit, a speed control unit, and a shaft torque control unit. In the one that calculates the torque current command and performs the electric inertia control via the inverter with the obtained torque current command,
The observer unit inputs a shaft torque detection value, a roller or dynamometer angular velocity detection value, calculates a roller surface driving force estimation value, and inputs a difference signal between the output of the observer and a running resistance value via the inertia unit. And a speed control controller for executing PID calculation with the roller or dynamometer angular velocity detection value, a deviation signal between the output of the speed controller and the shaft torque detection value, a dynamometer angular velocity detection value, and a roller angular velocity detection value A shaft torque controller that generates a torque current command by inputting a torque torque controller includes a shaft torque controller designed based on an ATR generalized plant model and designed by a controller design method called H∞ control and μ design method. An electrical inertia control device for a dynamometer system. The observer unit for calculating the roller surface driving force estimation value is constituted by a circuit designed by a controller design method called H∞ control and μ design method based on an observer generalized plant model. Item 2. An electric inertia control device for a dynamometer system according to Item 1.

Images